DESCRIPTION (adapted from Abstract): In the ACTG Trial 076 zidovudine (ZDV) administration during pregnancy and to infants during the first six weeks of life reduced significantly the rate of maternal-infant human immunodeficiency virus (HIV) transmission, the first intervention to do so. However, the high cost of ZDV therapy and the potential need for ZDV initiation early in pregnancy and intravenous administration during labor, make implementation of the full 076 protocol impractical in developing countries. Furthermore, the efficacy of this intervention in a population where breastfeeding is universal and considered essential for optimal nutrition and survival of infants needs to be determined. Therefore, the Principal Investigator proposes to evaluate two short-term ZDV regimens as means to reduce maternal-infant HIV transmission in a developing country setting (Haiti). One regimen will consist of maternal zidovudine beginning at 31-34 weeks gestation and continuing for two weeks postpartum, plus two weeks of therapy for infants; the other regimen will include maternal pre-, and intra- partum ZDV treatment but will not include postpartum treatment of the infant or the mother. Because ZDV is currently not available or affordable for routine use in Haiti, and because its potential toxicity in this population has not been assessed, the researchers will conduct a placebo- controlled trial evaluating the two short- course regimens in 940 mother- infant pairs. Women will receive 300 mg of ZDV twice a day during pregnancy and two weeks post-partum and 300 mg every three hours during labor. Infants randomized to receive ZDV will be given 2 mg/kg/dose four times per day for 14 days. Mothers and infants will be followed for 18 months; measured outcomes will include infection status and mortality. In addition, the association of maternal viremia with transmission of HIV and the effect of maternal use on the prevalence of HIV DNA, p24 antigen, and viable virus in breast milk will be evaluated.